Lafarge Construction Materials' standing orders call for a minimum mean freeboard of 18 in., which is also the barge's assigned load line. Based on the barge's verified deadweight scale, this freeboard equates to a total load (in salt water) of 3,551 t. Allowing for the local seawater density and the constant determined at the beginning of loading, this would have given a cargo load of 3,488 t. It was reported that the calculations made at 0345 indicated that a mean freeboard of 14 in. on completion would correspond to the amount of cargo the charterer had requested to be loaded aboard the barge. This freeboard would have equated to a total deadweight of some 3,644 t, and a final cargo figure of 3,595 t. Although the load line assigned to the SEASPAN195 did not apply at the time of the occurrence, at a freeboard of 14 in., the barge's draught exceeded her summer load line mark and that freeboard was less than the minimum specified in the charterer's Barge Loading Regulations. Although the weather was considered near optimum for the loading operation, the cargo dust suppression sprinkler system on the unloading boom was ineffective. Visibility on the tug and barge was severely restricted. Throughout the loading operation, the tug/barge master and mate were in communication by radiotelephone; however, due to the dust clouds caused by the loading, the master could not always see the distribution/disposition of the cargo from his vantage point aboard the tug. Further, as the actual loading rate (+/- 3000 metric tonnes/hour) was double that which had been agreed upon before loading of the barge commenced, it was difficult for the master to retain effective control of the distribution and trimming of the cargo. However, neither the master nor the mate requested that the loading rate be reduced. Information obtained during the investigation indicates that, with the exception of areas where loading continued while the barge was being shifted, most of the valleys between cargo peaks were not properly trimmed. Some of the peaks were considered to be higher than the previous norm. Because of the unequal areas which were clear and unused at the forward and after ends of the deck, it is apparent that distribution of the cargo on the barge was not symmetrical. After the draught survey, when the tug/barge deck-hand measured the after freeboards, the master, although reportedly concerned about the trim and amount of cargo yet to be loaded, did not intervene effectively to assist or advise the mate. Had some dialogue taken place between the master, the tug/barge mate and the ship's OOW, and a more precise loading sequence been formally agreed upon and maintained, the deck load could have been distributed so as to avert capsizing. Because the information given by the witnesses concerning the location, height and number of peaks of deck cargo varied, the distribution and related centres of gravity of the peaks could not be determined. Consequently, precise analysis of the loaded barge's stability characteristics shortly before the capsizing is precluded. Further, the determination of the actual cargo deadweight may not have been accurate because freeboard measurements were taken instead of a reading of the inherently more precise draught marks. However, it is known that the final runs of cargo were made to starboard on the sloping sides of the existing deck cargo peaks, and some cargo spilled over the top of the side wall of the cargo box. As the final loading continued, the resultant heeling effect caused the natural angle of repose of the sloping starboard sides of the existing cargo peaks to be exceeded, and initiated a sudden and accelerating athwartship shift of cargo. The combined effects of the shift of the cargo on the sloping starboard sides of the peaks and the continued final runs of cargo caused a heeling moment which exceeded the barge's righting ability, which resulted in the sudden capsizing to starboard.Analysis Lafarge Construction Materials' standing orders call for a minimum mean freeboard of 18 in., which is also the barge's assigned load line. Based on the barge's verified deadweight scale, this freeboard equates to a total load (in salt water) of 3,551 t. Allowing for the local seawater density and the constant determined at the beginning of loading, this would have given a cargo load of 3,488 t. It was reported that the calculations made at 0345 indicated that a mean freeboard of 14 in. on completion would correspond to the amount of cargo the charterer had requested to be loaded aboard the barge. This freeboard would have equated to a total deadweight of some 3,644 t, and a final cargo figure of 3,595 t. Although the load line assigned to the SEASPAN195 did not apply at the time of the occurrence, at a freeboard of 14 in., the barge's draught exceeded her summer load line mark and that freeboard was less than the minimum specified in the charterer's Barge Loading Regulations. Although the weather was considered near optimum for the loading operation, the cargo dust suppression sprinkler system on the unloading boom was ineffective. Visibility on the tug and barge was severely restricted. Throughout the loading operation, the tug/barge master and mate were in communication by radiotelephone; however, due to the dust clouds caused by the loading, the master could not always see the distribution/disposition of the cargo from his vantage point aboard the tug. Further, as the actual loading rate (+/- 3000 metric tonnes/hour) was double that which had been agreed upon before loading of the barge commenced, it was difficult for the master to retain effective control of the distribution and trimming of the cargo. However, neither the master nor the mate requested that the loading rate be reduced. Information obtained during the investigation indicates that, with the exception of areas where loading continued while the barge was being shifted, most of the valleys between cargo peaks were not properly trimmed. Some of the peaks were considered to be higher than the previous norm. Because of the unequal areas which were clear and unused at the forward and after ends of the deck, it is apparent that distribution of the cargo on the barge was not symmetrical. After the draught survey, when the tug/barge deck-hand measured the after freeboards, the master, although reportedly concerned about the trim and amount of cargo yet to be loaded, did not intervene effectively to assist or advise the mate. Had some dialogue taken place between the master, the tug/barge mate and the ship's OOW, and a more precise loading sequence been formally agreed upon and maintained, the deck load could have been distributed so as to avert capsizing. Because the information given by the witnesses concerning the location, height and number of peaks of deck cargo varied, the distribution and related centres of gravity of the peaks could not be determined. Consequently, precise analysis of the loaded barge's stability characteristics shortly before the capsizing is precluded. Further, the determination of the actual cargo deadweight may not have been accurate because freeboard measurements were taken instead of a reading of the inherently more precise draught marks. However, it is known that the final runs of cargo were made to starboard on the sloping sides of the existing deck cargo peaks, and some cargo spilled over the top of the side wall of the cargo box. As the final loading continued, the resultant heeling effect caused the natural angle of repose of the sloping starboard sides of the existing cargo peaks to be exceeded, and initiated a sudden and accelerating athwartship shift of cargo. The combined effects of the shift of the cargo on the sloping starboard sides of the peaks and the continued final runs of cargo caused a heeling moment which exceeded the barge's righting ability, which resulted in the sudden capsizing to starboard. The freeboard measurements taken immediately before the capsizing show that the SEASPAN195 was loaded with approximately 3,540 t of deck cargo. The deck cargo was not symmetrically distributed either transversely or longitudinally. The deck cargo was loaded in peaks of various heights, some of which were at a greater height than that to which the cargo was normally loaded. The final runs of cargo were made on the starboard side, the natural angle of repose of the sloping starboard sides of the existing cargo peaks was exceeded, some cargo shifted and spilled over the top of the side wall of the cargo box, and a sudden and accelerating athwartship shift of cargo occurred. The cumulative heeling and trimming moments caused by the shift of cargo to starboard overcame the barge's righting ability and initiated its capsizing. After the deck cargo self-discharged as the heel increased, the resultant buoyant upthrust on the lightened and nearly capsized barge caused it to rise in the water and become hooked by the port side wall of its cargo box on the main deck rails and gunwale on the port side of the ATLANTICSUPERIOR. The tug/barge mate suffered severe injuries to his legs when trapped between the barge and the deck of the ATLANTICSUPERIOR. Loading operation instructions were verbal; a written cargo plan, loading sequence, precise cargo deadweight and minimum finishing mean freeboards were not formalized before loading commenced. The assigned load line giving a minimum summer freeboard of 18 in. was not applicable to the service on which the barge was employed. The intended finishing mean freeboard of 14 in., as established from a draught survey carried out shortly before the capsizing, was less than the 18 in. minimum figure stipulated by the standing orders of the barge operators. The dust suppression water sprinkler system on the unloading boom did not prevent the formation of a dust cloud which impaired the ability of those involved in loading to see the distribution of the deck cargo. The adoption of a constant to account for an apparent increase in the verified lightship weight of the barge may have been the result of freeboard measurements being taken instead of a reading of the more precise draught marks. At the time of the occurrence, nothing in the company standing orders offered guidance to tug/barge crews on the loading or securing of barges while alongside self-unloaders.Findings The freeboard measurements taken immediately before the capsizing show that the SEASPAN195 was loaded with approximately 3,540 t of deck cargo. The deck cargo was not symmetrically distributed either transversely or longitudinally. The deck cargo was loaded in peaks of various heights, some of which were at a greater height than that to which the cargo was normally loaded. The final runs of cargo were made on the starboard side, the natural angle of repose of the sloping starboard sides of the existing cargo peaks was exceeded, some cargo shifted and spilled over the top of the side wall of the cargo box, and a sudden and accelerating athwartship shift of cargo occurred. The cumulative heeling and trimming moments caused by the shift of cargo to starboard overcame the barge's righting ability and initiated its capsizing. After the deck cargo self-discharged as the heel increased, the resultant buoyant upthrust on the lightened and nearly capsized barge caused it to rise in the water and become hooked by the port side wall of its cargo box on the main deck rails and gunwale on the port side of the ATLANTICSUPERIOR. The tug/barge mate suffered severe injuries to his legs when trapped between the barge and the deck of the ATLANTICSUPERIOR. Loading operation instructions were verbal; a written cargo plan, loading sequence, precise cargo deadweight and minimum finishing mean freeboards were not formalized before loading commenced. The assigned load line giving a minimum summer freeboard of 18 in. was not applicable to the service on which the barge was employed. The intended finishing mean freeboard of 14 in., as established from a draught survey carried out shortly before the capsizing, was less than the 18 in. minimum figure stipulated by the standing orders of the barge operators. The dust suppression water sprinkler system on the unloading boom did not prevent the formation of a dust cloud which impaired the ability of those involved in loading to see the distribution of the deck cargo. The adoption of a constant to account for an apparent increase in the verified lightship weight of the barge may have been the result of freeboard measurements being taken instead of a reading of the more precise draught marks. At the time of the occurrence, nothing in the company standing orders offered guidance to tug/barge crews on the loading or securing of barges while alongside self-unloaders. The SEASPAN195 capsized because unsymmetrical cargo distribution in conjunction with the heeling effects of the last runs of cargo initiated a shift of cargo on the sloping starboard sides of the existing deck cargo peaks, such that the combined heeling effects overcame the barge's righting ability. Visibility of the deck cargo from the JACQUES CARTIER was obscured by the dust cloud generated during loading operations which resulted in the creation of some unusually high peaks and contributed to the reported unsymmetrical distribution of the deck cargo.Causes and Contributing Factors The SEASPAN195 capsized because unsymmetrical cargo distribution in conjunction with the heeling effects of the last runs of cargo initiated a shift of cargo on the sloping starboard sides of the existing deck cargo peaks, such that the combined heeling effects overcame the barge's righting ability. Visibility of the deck cargo from the JACQUES CARTIER was obscured by the dust cloud generated during loading operations which resulted in the creation of some unusually high peaks and contributed to the reported unsymmetrical distribution of the deck cargo.